Conferences

ESRA2009:

Conference main page

Overview of sessions

Time table

Warsaw 2009: Presentations and short courses

Three approaches to analyzing panel data: Latent growth curve models, autoregressive models and stochastic differential equations

Session: Comparing and Evaluating Autoregressive, Latent Trajectory, Autoregressive Latent Trajectory, and Continuous Time ALT Models (II)

Authors:

Peter Schmidt; University of Giessen, Germany
Eldad Davidov; University of Zürich, Switzerland
Han Oud; Radboud Universiteit Nijmegen, Netherlands

Abstract:

Autoregressive (AR) models and latent growth curve (LGC) models are two well-established methods for the analysis of panel data. In the past, much research has focused on the comparison of the two approaches (e.g., Raykov, 1998), whereas the idea of combining them – so called autoregressive latent trajectory (ALT) models – is comparatively new (Bollen & Curran, 2004). For
all three models, however, data are assessed at discrete time points and little is known about the robustness of the findings if time points would have been different. The present paper takes up this issue and shows how differential equation modeling provides a solution to this problem by
linking the discrete time model parameters to the underlying continuous time model. The different approaches are compared and illustrated using data from the Group-Focused Enmity Survey, which was conducted at three time points between 2002 and 2004.